Spectral analysis of respiratory-related hemodynamic variables in simulated hypovolemia: a study in healthy volunteers with spontaneous breathing using a paced breathing activity

BACKGROUND A dynamic preload index such as stroke volume variation (SVV) is not as reliable in spontaneous breathing (SB) patients as in mechanically ventilated patients. This study examined the hypothesis that spectral analysis of hemodynamic variables during paced breathing (PB) activity may be a feasible index of volume changes and fluid responsiveness, despite insufficient respiratory changes in the preload index during SB activity. METHODS Blood pressure and stroke volume (SV) were measured in 16 subjects undergoing PB (15 breaths/min), using a Finometer device and the Modelflow method. Respiratory systolic pressure variation (SPV) and SVV were measured and respiratory frequency (RF, 0.2-0.3 Hz) of power spectra of SPV (SPV(RF)) and SVV (SVV(RF)) were computed using fast Fourier transformation. Progressive hypovolemia was simulated with lower body negative pressure (LBNP). Volume challenges were produced by infusion of normal saline and subsequent release of LBNP to baseline. Fluid responsiveness, defined as a >20% increase in SV, was assessed by the area under the curve (AUC) of receiver operating characteristic curves. RESULTS Graded hypovolemia caused a significant increase in SPV(RF) and a decrease in SVV(RF). During volume expansion, SPV(RF) decreased and SVV(RF) rose significantly. Fluid responsiveness was better predicted with SVV(RF) (AUC 0.75) than with SPV(RF), SPV, or SVV. SVV(RF) before volume challenge was significantly correlated with volume expansion-induced changes in SV (r = -0.64). CONCLUSIONS These results suggest that RF spectral analysis of dynamic preload variables may enable the detection of volume change and fluid responsiveness in SB hypovolemic patients performing PB activity.